Examples with KV org.apache.beam.sdk.values.KV used on opensource projects

Example 31 with KV

use of org.apache.beam.sdk.values.KV in project beam by apache.

the class ParDoTest method testSetStateCoderInferenceFailure.

@Test
@Category({ ValidatesRunner.class, UsesStatefulParDo.class, UsesSetState.class })
public void testSetStateCoderInferenceFailure() throws Exception {
    final String stateId = "foo";
    final String countStateId = "count";
    Coder<MyInteger> myIntegerCoder = MyIntegerCoder.of();
    DoFn<KV<String, Integer>, Set<MyInteger>> fn = new DoFn<KV<String, Integer>, Set<MyInteger>>() {

        @StateId(stateId)
        private final StateSpec<SetState<MyInteger>> setState = StateSpecs.set();

        @StateId(countStateId)
        private final StateSpec<CombiningState<Integer, int[], Integer>> countState = StateSpecs.combiningFromInputInternal(VarIntCoder.of(), Sum.ofIntegers());

        @ProcessElement
        public void processElement(ProcessContext c, @StateId(stateId) SetState<MyInteger> state, @StateId(countStateId) CombiningState<Integer, int[], Integer> count) {
            state.add(new MyInteger(c.element().getValue()));
            count.add(1);
            if (count.read() >= 4) {
                Set<MyInteger> set = Sets.newHashSet(state.read());
                c.output(set);
            }
        }
    };
    thrown.expect(RuntimeException.class);
    thrown.expectMessage("Unable to infer a coder for SetState and no Coder was specified.");
    pipeline.apply(Create.of(KV.of("hello", 97), KV.of("hello", 42), KV.of("hello", 42), KV.of("hello", 12))).apply(ParDo.of(fn)).setCoder(SetCoder.of(myIntegerCoder));
    pipeline.run();
}

Example 32 with KV

use of org.apache.beam.sdk.values.KV in project beam by apache.

the class ParDoTest method testMapStateCoderInference.

@Test
@Category({ ValidatesRunner.class, UsesStatefulParDo.class, UsesMapState.class })
public void testMapStateCoderInference() {
    final String stateId = "foo";
    final String countStateId = "count";
    Coder<MyInteger> myIntegerCoder = MyIntegerCoder.of();
    pipeline.getCoderRegistry().registerCoderForClass(MyInteger.class, myIntegerCoder);
    DoFn<KV<String, KV<String, Integer>>, KV<String, MyInteger>> fn = new DoFn<KV<String, KV<String, Integer>>, KV<String, MyInteger>>() {

        @StateId(stateId)
        private final StateSpec<MapState<String, MyInteger>> mapState = StateSpecs.map();

        @StateId(countStateId)
        private final StateSpec<CombiningState<Integer, int[], Integer>> countState = StateSpecs.combiningFromInputInternal(VarIntCoder.of(), Sum.ofIntegers());

        @ProcessElement
        public void processElement(ProcessContext c, @StateId(stateId) MapState<String, MyInteger> state, @StateId(countStateId) CombiningState<Integer, int[], Integer> count) {
            KV<String, Integer> value = c.element().getValue();
            state.put(value.getKey(), new MyInteger(value.getValue()));
            count.add(1);
            if (count.read() >= 4) {
                Iterable<Map.Entry<String, MyInteger>> iterate = state.entries().read();
                for (Map.Entry<String, MyInteger> entry : iterate) {
                    c.output(KV.of(entry.getKey(), entry.getValue()));
                }
            }
        }
    };
    PCollection<KV<String, MyInteger>> output = pipeline.apply(Create.of(KV.of("hello", KV.of("a", 97)), KV.of("hello", KV.of("b", 42)), KV.of("hello", KV.of("b", 42)), KV.of("hello", KV.of("c", 12)))).apply(ParDo.of(fn)).setCoder(KvCoder.of(StringUtf8Coder.of(), myIntegerCoder));
    PAssert.that(output).containsInAnyOrder(KV.of("a", new MyInteger(97)), KV.of("b", new MyInteger(42)), KV.of("c", new MyInteger(12)));
    pipeline.run();
}

Example 33 with KV

use of org.apache.beam.sdk.values.KV in project beam by apache.

the class ParDoTest method testCombiningStateCoderInference.

@Test
@Category({ ValidatesRunner.class, UsesStatefulParDo.class })
public void testCombiningStateCoderInference() {
    pipeline.getCoderRegistry().registerCoderForClass(MyInteger.class, MyIntegerCoder.of());
    final String stateId = "foo";
    DoFn<KV<String, Integer>, String> fn = new DoFn<KV<String, Integer>, String>() {

        private static final int EXPECTED_SUM = 16;

        @StateId(stateId)
        private final StateSpec<CombiningState<Integer, MyInteger, Integer>> combiningState = StateSpecs.combining(new Combine.CombineFn<Integer, MyInteger, Integer>() {

            @Override
            public MyInteger createAccumulator() {
                return new MyInteger(0);
            }

            @Override
            public MyInteger addInput(MyInteger accumulator, Integer input) {
                return new MyInteger(accumulator.getValue() + input);
            }

            @Override
            public MyInteger mergeAccumulators(Iterable<MyInteger> accumulators) {
                int newValue = 0;
                for (MyInteger myInteger : accumulators) {
                    newValue += myInteger.getValue();
                }
                return new MyInteger(newValue);
            }

            @Override
            public Integer extractOutput(MyInteger accumulator) {
                return accumulator.getValue();
            }
        });

        @ProcessElement
        public void processElement(ProcessContext c, @StateId(stateId) CombiningState<Integer, MyInteger, Integer> state) {
            state.add(c.element().getValue());
            Integer currentValue = state.read();
            if (currentValue == EXPECTED_SUM) {
                c.output("right on");
            }
        }
    };
    PCollection<String> output = pipeline.apply(Create.of(KV.of("hello", 3), KV.of("hello", 6), KV.of("hello", 7))).apply(ParDo.of(fn));
    // There should only be one moment at which the average is exactly 16
    PAssert.that(output).containsInAnyOrder("right on");
    pipeline.run();
}

Example 34 with KV

use of org.apache.beam.sdk.values.KV in project beam by apache.

the class ParDoTest method testOutOfBoundsEventTimeTimer.

@Test
@Category({ ValidatesRunner.class, UsesTimersInParDo.class })
public void testOutOfBoundsEventTimeTimer() throws Exception {
    final String timerId = "foo";
    DoFn<KV<String, Integer>, Integer> fn = new DoFn<KV<String, Integer>, Integer>() {

        @TimerId(timerId)
        private final TimerSpec spec = TimerSpecs.timer(TimeDomain.EVENT_TIME);

        @ProcessElement
        public void processElement(ProcessContext context, BoundedWindow window, @TimerId(timerId) Timer timer) {
            timer.set(window.maxTimestamp().plus(1L));
        }

        @OnTimer(timerId)
        public void onTimer(OnTimerContext context) {
        }
    };
    PCollection<Integer> output = pipeline.apply(Create.of(KV.of("hello", 37))).apply(ParDo.of(fn));
    thrown.expect(RuntimeException.class);
    // Note that runners can reasonably vary their message - this matcher should be flexible
    // and can be evolved.
    thrown.expectMessage("event time timer");
    thrown.expectMessage("expiration");
    pipeline.run();
}

Example 35 with KV

use of org.apache.beam.sdk.values.KV in project beam by apache.

the class ParDoTest method testValueStateCoderInference.

@Test
@Category({ ValidatesRunner.class, UsesStatefulParDo.class })
public void testValueStateCoderInference() {
    final String stateId = "foo";
    MyIntegerCoder myIntegerCoder = MyIntegerCoder.of();
    pipeline.getCoderRegistry().registerCoderForClass(MyInteger.class, myIntegerCoder);
    DoFn<KV<String, Integer>, MyInteger> fn = new DoFn<KV<String, Integer>, MyInteger>() {

        @StateId(stateId)
        private final StateSpec<ValueState<MyInteger>> intState = StateSpecs.value();

        @ProcessElement
        public void processElement(ProcessContext c, @StateId(stateId) ValueState<MyInteger> state) {
            MyInteger currentValue = MoreObjects.firstNonNull(state.read(), new MyInteger(0));
            c.output(currentValue);
            state.write(new MyInteger(currentValue.getValue() + 1));
        }
    };
    PCollection<MyInteger> output = pipeline.apply(Create.of(KV.of("hello", 42), KV.of("hello", 97), KV.of("hello", 84))).apply(ParDo.of(fn)).setCoder(myIntegerCoder);
    PAssert.that(output).containsInAnyOrder(new MyInteger(0), new MyInteger(1), new MyInteger(2));
    pipeline.run();
}